Our guest this week is Michael Shermer, the publish of Skeptic magazine and head of the Skeptics Society, and a longtime commentator on issues relating to science, critical thinking, and the paranormal.
Chris asked Michael on to discuss his new book, which is entitled The Believing Brain: From Ghosts and Gods to Politics and Conspiracies, How We Construct Beliefs and Reinforce Them As Truths.
Clearly, much of what Shermer has to say here will be of great relevance to skeptics and freethinkers—and along the way, Shermer also discusses his views on global warming (real, but not such a big deal) and how to promote evolution in a religious America.
In addition to publishing Skeptic, Michael Shermer is a monthly columnist for Scientific American, the host of the Skeptics Distinguished Science Lecture Series at Caltech, and Adjunct Professor at Claremont Graduate University. His other books include Why People Believe in Weird Things and Why Darwin Matters.
Interesting bit about global warming that I wanted to respond to, even though I’m not done listening yet. Even though he believes it, Shermer doesn’t seem worried about it. I’ve kind of been feeling the same way, which may explain why I’m not really bothered to debate AGW skeptics. Since the consequences of global warming are on the timescale of decades the effects of technological advance must be taken into account, and if Ray Kurzweil is even half-right in his predictions we should expect the pace of technological evolution to continue to accelerate, and considering the new focus on green technology, I feel we as a species are likely to come up with technological solutions before the consequences of global warming become too great. Maybe Shermer was not thinking along these lines, but I think many people, scientists and politicians included, don’t recognize the acceleration of the rate of technological advance and may not be taking it into account when they raise their alarm bells.
If we reach human-level artificial intelligence within the century (Kurzweil predicts 2045), it could be in time to have an impact on the consequences of global warming. But of course, lesser technologies will also have an effect.
History has shown Kurzweil’s predictions to be on the optimistic side (in terms of time horizon), but not by a large amount. The singularity-as-rapture stuff I don’t really buy, but I don’t see how one can rationally deny that machines are getting smarter at an accelerating pace.
Technological advance as a whole does seem to be accelerating for whatever reason, be it new technologies creating more innovation or population growth creating more inventors, or a mixture of the two. I agree we should take steps towards addressing global warming now, but I also think the future will provide new solutions. Look at the human genome project. They thought it would take much longer than it did, but thanks to technological progress it was completed way ahead of schedule.
History has shown Kurzweil’s predictions to be on the optimistic side (in terms of time horizon), but not by a large amount. The singularity-as-rapture stuff I don’t really buy, but I don’t see how one can rationally deny that machines are getting smarter at an accelerating pace.
How are machines getting smarter? Faster, yes, but smarter is more nebulous. Sure, computers can beat humans at chess and Jeopardy ®, but those tasks require processing power, not intelligence. I’ll believe computers are smarter than humans when one spits out a Unified Field Theory.
As for Kurzweil, his correct predictions were stuff many others also predicted. I’m not impressed that he claims to have foreseen portable computing devices. I foresaw those too, as did the people in the computer industry who were designing the devices.
I don’t know that there’s a principled distinction to be made between processing power and intelligence, at least not when processing power is being put to tasks that involve learning, memory and reason. What I would say is that it’s clear increasing processing power is a relatively easy task to accomplish, when compared with increasing machine intelligence.
I’d agree with domokato that machines are getting smarter (they are able to perform more complex reasoning tasks), and that the pace of increase may be accelerating. But that’s not to say Kurzweil’s predictions are at all sensible.
Since the consequences of global warming are on the timescale of decades the effects of technological advance must be taken into account, and if Ray Kurzweil is even half-right in his predictions we should expect the pace of technological evolution to continue to accelerate, and considering the new focus on green technology, I feel we as a species are likely to come up with technological solutions before the consequences of global warming become too great.
AGW is accelerating too. It would go on even if we would stop producing CO2 now. But instead we are only reducing the acceleration of producing CO2. You want to bet, with our species as stake, what is faster? Technological evolution or AGW?
I don’t know that there’s a principled distinction to be made between processing power and intelligence, at least not when processing power is being put to tasks that involve learning, memory and reason. What I would say is that it’s clear increasing processing power is a relatively easy task to accomplish, when compared with increasing machine intelligence.
I don’t understand this. First you say that processing power and intelligence are not that different, with which I agree, but then you go on to say that one is easier to accomplish than the other. So are they different or not?
How are machines getting smarter? Faster, yes, but smarter is more nebulous.
I believe computers are getting smarter. Photoshop from ten years ago had only one undo, now that number is much higher. The illusion that one may “decide” to come up with a Unified Field Theory doesn’t really differ from Photoshop “being told” by a programmer to allow for multiple number of undos. I know it smells like another Free Will topic, but that’s the way things usually turn out when it comes to this kind of topics.
And if you don’t like calling today’s Photoshop smarter than Photoshop from ten years ago it may also be a mistake to call a person who may figure out a Unified Field Theory smarter than a burger flipper at McDonald’s. A burger flipper might not have the necessary brain connections to figure out a Unified Field Theory just like Photoshop from ten years ago didn’t have the necessary connections to perform more than one undo.
I don’t know that there’s a principled distinction to be made between processing power and intelligence, at least not when processing power is being put to tasks that involve learning, memory and reason. What I would say is that it’s clear increasing processing power is a relatively easy task to accomplish, when compared with increasing machine intelligence.
I don’t understand this. First you say that processing power and intelligence are not that different, with which I agree, but then you go on to say that one is easier to accomplish than the other. So are they different or not?
Well, it’s like stones and a house. A house is made of stones, but a house is more than just a pile of stones. Intelligence is made of processing power but intelligence is more than just raw processing power. It’s processing power organized in a certain way.
So I take it that in your opinion there is a principled distinction to be made between processing power and intelligence (?).
There isn’t “when processing power is being put to tasks that involve learning, memory and reason”, as I said before. Raw processing power isn’t the same as intelligence. Processing power used to learn, remember and reason is intelligence.
How are machines getting smarter? Faster, yes, but smarter is more nebulous.
I believe computers are getting smarter. Photoshop from ten years ago had only one undo, now that number is much higher.
And if you don’t like calling today’s Photoshop smarter than Photoshop from ten years ago it may also be a mistake to call a person who may figure out a Unified Field Theory smarter than a burger flipper at McDonald’s. A burger flipper might not have the necessary brain connections to figure out a Unified Field Theory just like Photoshop from ten years ago didn’t have the necessary connections to perform more than one undo.
Multiple undos could be 90% the same code as single undos just with recursion. That could be a matter of hard disk drive size. A 500 gigabyte drive today is the same price as a 20 gig drive back then and 500 gig drives did not exist then as far as I know.
I would just say faster and with more capacity not smarter and most comparisons of computers to the human brain/mind are nonsense. von Neumann machines manipulate symbols in the form of bit combinations. They do not comprehend what the bit combinations mean.
Can you load a picture of a herd of cows into Photoshop and have the computer explain what the picture is? Do you expect 5 year old kid to be able to do that?
So I take it that in your opinion there is a principled distinction to be made between processing power and intelligence (?).
There isn’t “when processing power is being put to tasks that involve learning, memory and reason”, as I said before. Raw processing power isn’t the same as intelligence. Processing power used to learn, remember and reason is intelligence.
Are you saying that computers are not intelligent (or not easily made to be intelligent) because they don’t (easily) learn, remember and reason?
